Ellen, Gretchen, and Lisa, not yet 3, were locked for many weeks in a social triangle, wrenched by escalating clashes over who got to play with whom at any given time. Like other young preschool girls, each one strongly preferred to play in a dyad (Benenson 1993). As all three initially refused to play with any other children, one of the girls was always left out. The excluded child would invariably throw a tantrum, screaming at and sometimes even physically attacking the other two. Even more upsetting, the trio rearranged their pairings several times an hour. Each shift would result in tears, shouts, frustration, and anger. This deeply stressed the girls, interrupted their play time, and sharply disturbed the other children’s activities. As scenes such as those unfolded, I asked myself: What micro-events, what confluence of factors, causes these abrupt re-pairings? If I could answer that, I might be able to help ease them past their conflicts less painfully.

As a veteran early childhood educator, I don’t expect a conflict-free classroom. But these were not the typical toddler disputes. They were similar to the pitched emotional battles over friendships and play partners often seen in older children (Underwood 2003). Oddly, too, this was the second year in a row my class included a very contentious play triangle. The previous time it had been three boys, so gender didn’t seem to be a factor. The children’s harsh strife, deep anguish at being excluded, and disruption of other children’s play all greatly troubled me. This led me to the questions at the heart of my teacher research: How could my coteachers and I figure out what was going on in the girls’ interactions? What could we do to help these children manage their relationships? My initial impulse was to somehow understand the source of these painful incidents, mainly so we might anticipate and prevent them.

Review of literature

A large body of literature describes typical child development and the evolving cognitive and affective capabilities that underlie young children’s growing social competence. Much research describes the typical maturation of emotion, empathy, thinking, and language involved in play among 2s and 3s (Brownell & Kopp 2007). However, practical help for educators regarding classroom management strategies, and teachers’ roles in scaffolding positive peer interactions with older toddlers, is limited within early childhood research. Its inquiries usually focus on correlates and effects of behavior. A few studies describe teacher strategies to help less-skilled children enter play with available partners (Hazen & Black 1984; Bullock 1988; Fox & Lentini 2006; Gillespie & Hunter 2010), although this wasn’t quite the situation we faced. The same was true of other studies that examine how children interact in short-lived, three-way child encounters in the laboratory (McLloyd et al. 1984; Ishikawa & Hay 2006). However, I could find no naturalistic or clinical research about young preschoolers involved in exclusive, persistent, and highly conflicting groups of three.

Peer-culture studies helped me better understand the trio’s tumultuous encounters (Kantor et al. 1993; Corsaro 2003; Neimark 2012). As I read this work I began to see that children’s exclusions and rejections were not necessarily cruel or even artless. The central tenet of such studies is that children are social agents, shaping their relational networks and seeking to control their social worlds outside of adult agendas. The peer culture literature reminds teachers to focus on the child’s point of view, on how any given behavior is meant to serve her goals of creating knowledge or connection (Avgitidou 2001). In this view, when children hoard certain toys, dominate certain spaces at school, invent private language or rituals, speak harshly to each other, or exclude each other (all behavior that Ellen, Lisa, and Gretchen used regularly), these interactions serve as little workshops in which they gradually forge satisfying processes of play and companionship.

Older toddlers, more aware of the delights of shared play and gradually moving into more collaborative play, are just learning the complicated skills of keeping these play connections going. Their play with other children may feel so fragile and precarious to them (Corsaro 2003) that they act to protect it with exclusionary behavior or verbal aggression. Such acts may not be the evidence of selfishness, defiance, or even antisocial impulses that many adults tend to assume. As I started this study, the intensity of the triad’s conflicts, and their appearance so early in preschool, tended to make me wonder if the girls’ hurtful behavior resulted from emotional disequilibrium or even aggressive traits.

The constructivist view of child development says that children actively assemble knowledge and meaning, rather than just imitating others or passively receiving understandings. Piaget (1966) saw this process of cognitive growth as a mainly internal, solitary process, with only peripheral attention to the role of play with others. Vygotsky (1978), in contrast, described such learning as occurring mainly in social contexts, during moments of close collaboration in shared activities. The notion that children develop an ability to see that other people have feelings or understandings different than their own is called the “theory of mind.” This ability is key to collaborative learning. How were the three girls learning from each other’s reactions in these situations, and what were they learning?

The sociocultural perspective of Rogoff and others provides another frame for this teacher research. The customary practices, values, and language of a society shape its children’s encounters with each other. Children’s behavior mirrors cultural norms of emotional expression, friendship, community, status, conflict, and property (Rogoff 2003). Culture and economic class also mold the teaching methods and child behavior that early childhood teachers consider appropriate (Tobin et al. 2011). For instance, in Japanese culture, preschool teachers permit other children to tease or exclude certain girls so as to evoke their amae, or culturally valuable feeling of loneliness, and sudan shugi, or social mindedness (Hayashi & Tobin 2009). My own cultural model led me to suppress conflicts and “negative” feelings.

My frame of reference

My upbringing provided another frame, teaching me to see the exclusion of others, or the refusal to share, as selfish and egoistic. To nurture children and better our society, I’d always taught children to treat each other gently and avoid excluding others as part of building their sense of community. Prior to this study, in my role as a teacher I uncritically endorsed the “you can’t say you can’t play” guidelines (Paley 1992; Harrist & Bradley 2003), automatically insisting in each situation that children include all others who wish to join in.

My own teacher training also affected my thinking about conflict between children. Echoing Piaget’s theories, it taught me that young children are egoistic thinkers, and therefore, empathetic behavior couldn’t develop until later in childhood (Piaget 1966). I learned from this that young children require adult instruction to form positive social attitudes, and that children’s clashes were to be minimized as antisocial.

My initial research on the sources of the girls’ painful interactions challenged prior assumptions. Looking closely into the causation of the girls’ sudden regroupings, exploring what insight they gained and how they gleaned social knowledge, all changed my perspective. I had wanted to see how my colleagues and I could lessen classroom discord and help children manage their relationships. What I didn’t know beforehand was that my study would greatly shift how I understood the functions of conflict, which in turn influenced how the children related with each other.

Design of the study

The participants and the setting

My teacher research project included three children and three teachers in the oldest toddler class of a private preschool. I started collecting data halfway through the school year. Lisa, who was sturdy, lively, passionate, and opinionated, was most likely to express negative feelings about the others’ behavior, and often tried to direct their play. Ellen, verbal and impish, often negotiated her way back into a dyad by using skills of planning and persuasion. Gretchen was the most outwardly emotional one, most upset by exclusions or frustrations. She needed the most help from teachers to regain her equilibrium. All three loved dancing, chasing, trying new materials at the art table, and playing under the climbing structure or in the dramatic play area.

The children came from intact, two-parent, middle-class families. All were Caucasian, like 76 percent of my students. The 17 children in the class ranged in age from 24 to 34 months when they started in the fall.

My school is a private nonprofit in a university town with a full-day, play-based program using an emergent curriculum approach and emphasizing social and emotional learning. Our class, housed in a modest-sized room, has many learning materials available on open shelves and an inviting dramatic play area in a roomy alcove three steps up from the main room. We have our own yard, surfaced with sand and lined by foliage, with a redwood tree in the corner, a playhouse, a climbing structure, and a dome. The teachers were Martha, Annie, and myself. Martha and Annie each had more than three years of teaching experience; I have taught for more than 30 years.

Research plan, data collection, and analysis

As the study began, I avoided intervening in the girls’ conflicts because I intended to study how they dealt with each other free of my interference. However, I watched to make sure they weren’t harming each other, stepping in to guide them only when they became too overwrought to cope. Yet at the same time, from my own experience I knew that very young children also need extra help in relating. As one teacher researcher wrote, “I realize more strongly now the importance of building and guiding friendships for toddlers. I need[ed] to learn the different cues that toddlers use to [communicate to understand when and what kind of intervention might] promote friendship” (Meier & Henderson 2007, 104). As the study progressed, therefore, our teaching team decided to offer more active assistance.

Data collection

In order to answer my questions, over a period of five months I documented the children’s interactions with each other throughout the day, recording my interpretations and those of the teachers and parents. I began my research by creating various time charts to log the girls’ interactions by the minute and get a clearer view of their social dynamics. This proved unworkable because I had little chance to write many notes and because encounters didn’t fit into neat five-minute time slots. I then created a simpler, split-page chart on which I could record the social “moves” that accompanied their re-pairings. My data included transcripts of the girls’ conversations, reflections in my journal, a few photos, and the field notes about their interactions. I also used my coteachers’ observations, checking my field notes and analyses with them to strengthen accuracy. I spoke about the children’s behavior at home with their parents, who described the usual sibling conflicts and the mostly typical behavior of 3-year-olds. I separately interviewed each girl (once in February, once in April) about her likes and dislikes, her playmate preferences, and her feelings about the others.

Analysis methodology

I analyzed my field notes and transcripts and identified anecdotes that illustrated key social behavior and teacher intervention methods. I coded them by date, type of interaction, participants, emotional tone, the resulting dyad after conflict (or cooperation), and the language of inclusion and exclusion they used. I looked for patterns, beginning with the order of the three girls’ arrival at school each morning as they arranged their initial dyad formation. I also considered how personality and temperament might be shaping their interactions.

Findings and analysis

This study helped me construct a more detailed understanding of how children learn about social connections and dynamics through their experiments with exclusion, pairing-off, verbal exchanges, and hurt feelings. I was able to see some effects of this new understanding on the children and on my classroom practice. My first finding was that conflict among children is not always negative, requiring adult intervention. This called my previous beliefs into question. The data I collected about these highly interactive conflicts challenged the widely held conclusion that young children are chiefly egoistic, inherently insensitive to each other, and lacking in empathy or social awareness. The three children’s conflicts were real, and often hurtful, which is why initially I saw them primarily from a disapproving adult perspective.

My data reinforced the finding that peer cultures among preschoolers involve a resistive stance toward adult expectations and rules that teachers often mistake as defiant or headstrong (Corsaro 2003). But children creating peer cultures bend adult rules more out of a desire for self-identity and creativity, to mark them as the children’s inventions, slightly outside of adult control. The triad’s whole pattern of noisy conflict fit this feature of peer cultures. The girls’ frequent re-pairings gave each one many chances to violate teacher expectations in a variety of roles.

Through analyzing the data, I came to see that deficit-based thinking about children’s behavior during conflict only obscures useful insights about such behavior, and thus undermines effective teaching. Instead, I began to understand that these girls exhibited much more sophisticated social skills at this early age than I had thought possible. I was often jolted by the girls’ persuasive abilities and their maneuvers to shape their world of play relationships, as in this excerpt.

From Field Notes, February 5:

The children’s display of interpersonal awareness and planning surprised and impressed me with its intricacy and intentionality. I felt that Ellen arranged this chance to be alone with Gretchen (a captive audience) to monopolize her loyalty and seek Gretchen’s vow not to play with Lisa. Gretchen at first seemed to evade the issue, but wound up fishing for Ellen’s statement of commitment to her (“Are you Lisa’s friend?”). Ellen pressed on, telling Gretchen that she was her preferred playmate. Moment by moment, both girls seemed to be assessing the other’s words and intonations, intuiting each other’s intentions, seeking to modify each other’s stance, and adjusting their maneuvers as they engaged in purposeful planning to exclude the third girl.

In looking at the patterns across my data, I found that each girl was aware of her own emotional states and grew more aware of the feelings of the other two. Each deeply felt the pain of exclusion or the triumph of being “chosen.” However, even though each girl seemed aware of the grief of being the “odd man out,” she nevertheless always chose the privilege to play as a dyad when it was offered—or when she could arrange to get it.

Another finding suggested by the data was that the girls were perceptive observers. Their jockeying for dyad connection showed how they paid attention to each other—a cognitive as well as affective advance. They derived interpersonal knowledge from facial expressions and postures. They seemed able to predict each other’s behavior, plan future action, intuit each other’s states of mind, and act to influence, delay gratification, and even assess each other’s play potential (Corsaro 2003; Neimark 2012). Each girl created tactics to get herself “in good with” the desired partner, using knowledge of her play preferences or family ties. For example, in the February 5 field notes, Ellen shifted the awkward conversation by mentioning the older brother whom Gretchen adored.

The girls also appeared to monitor their own efforts to influence others’ actions and adjusted their strategies as needed—another example of their increased skills of self-regulation. For example, on April 15, as Lisa and Gretchen lay giggling beneath the favorite quilt on the play bed, Ellen playfully leaped on top of them and begged Lisa to come with her. Lisa told her to go away. Using her awareness that Lisa loved baby care, Ellen switched tactics. She snatched up a doll and crooned temptingly, “Here’s your baby, Lisa, here’s your baby! C’mon!” Considering Lisa’s feelings enabled Ellen to choose an other-centered strategy—though in the end Lisa didn’t go off with her as she had hoped!

Teacher approaches

As the study progressed, my approach to the children was modified by my enhanced understanding of their exchanges. I came to see that tussling over their alignments allowed them to explore interpersonal power and negotiate social mazes. My early response to the triad conflicts had been automatic rather than thoughtful. Initially, when conflicts occurred I found that my efforts to lecture or to suppress the conflict didn’t work, as the following vignette shows.

We teachers conferred privately to find more effective responses. As adults we couldn’t “make over” their play arrangements, much less guarantee their tranquility. Instead, to model gentler language and sociable behavior, we set out to role-play some different ways to interact. On four occasions my coteacher Martha staged short plays for all the children at story time (using puppets) in which the characters excluded each other and acted upset. Then she asked the class to suggest other ways the characters could speak to each other. She helped the children develop some new responses, such as “You can play with us later!” instead of “You can’t play with us!” Martha then reenacted the scene using these new phrases, which yielded happier outcomes for the puppets.

Suggesting other playmates

By February 20, we’d seen the girls occasionally try playing with girls outside of the triad. A few days later, I noted first Gretchen, then Lisa, become engaged with Natasha. After spring break we teachers started a new strategy. One day when the other two noisily excluded Gretchen, I waited a bit, and then murmured to her that Charlene might like to play. They clicked! I was surprised, since the trio had seemed so ingrown and the newer playmates were developmentally younger. I noticed, however, that the alternate playmates got abandoned in a flash whenever a triad member became available for play. These ventures into wider play choices prompted us to urge the girls’ parents to set up playdates with various children outside the triad.

Broadening choices with “friendly” speech

Despite the play with new friends, the three girls kept up their taunting language toward each other. So in mid-March, we teachers invented a game called “Friendly Words” to foster kindness in social speech (Doescher & Sugarawa 1992). Since tone conveys meaning, the teachers took turns uttering common phrases in a variety of tones, then asking, “Are those words ‘friendly’ or ‘unfriendly’?” We varied the order of the choices to avoid cuing the children, and we implied no right or wrong answers. We didn’t say “nice”/“not nice” because these loaded phrases are commonly used by adults to evaluate children themselves. We might, for example, say “I don’t want to play right now. Maybe later,” in a pleasant tone. Some children called that “unfriendly,” others “friendly.” Some could even explain why they “voted” as they did, revealing their skills in perception, causation, and language. In following days, we also asked children at random moments to rate classroom exchanges that we overheard as “friendly” or “unfriendly.” Soon I heard more new snippets of conversation.

Lisa’s answer itself might have been in part a performance, another form of the rule-breaking peer culture that expresses children’s independence of adult expectations.

Here’s an example of how the three girls began to develop more nuanced social skills.

In this interaction, the three girls diversified their choices of social discourse. Their exchanges back in January had often seemed angrier, reactive—revealing little ability to wait for the favored playmate or to refocus their attention when disappointed. Earlier in the year, Lisa would never have ignored Ellen to keep playing with Charlene, and Lisa and Ellen wouldn’t have discussed their divergent plans so calmly. In this exchange Ellen may have felt angry with Lisa (“I’m going to play with you later”), but her answer to Lisa’s regretful protest (“Do you have to, Ellen?”) sounded poignant, not angry: “You hurt my feelings.” Lisa’s response feels like a gentle attempt to reconcile with Ellen. The growing frequency of their consideration and negotiation were not just due to our coaching, of course, but were based also in new choices they were making about relating.

My notes show that talk between the three girls always accompanied (or sharpened) their conflicts, unlike with younger toddlers who clash using fewer words. Their language of inclusion and exclusion was so ritualized that I learned to abbreviate their phrases as letter strings (for example, YCBMYF was “You can’t be my friend”). Over time they began to echo the teachers’ models of social language (“I’ll play with you later” and “You hurt my feelings”) and to develop their own versions (“Will you play with me when we go outside?”). Also, they learned to connect more successfully by using language to fit into the existing dynamics of the play or to suggest a new game (“Want to take your boots off? Get dressed up?”).

My field notes showed that by mid-April, intense conflicts decreased in frequency as the girls chose gentler interactions. The girls sometimes played with “outsiders,” used more friendly tones, and joined in play as a trio for brief periods. Also, I saw times when the “out” girl approached the dyad and the other two would gently redirect her to other play, instead of just taunting or baldly rejecting her.

Another possible finding for the changed behavior is that each of the three grew to care more deeply for the other two. Data from late in the study showed the girls striving at times to meet each other’s needs and to avoid mutual harm. On April 7, Gretchen went to sit with Ellen, Lisa, and teacher Martha at the picnic table but began crying that they hadn’t left room. Ellen swiftly scooted over to allow Gretchen to sit beside her. Their mutual caring may have come to seem as valuable to each one as the temporary pleasure of having the dyad partner all to herself.

Conclusions and implications

Through my study, I came to understand how my preconceptions about children’s conflicts had diminished my teaching success; formerly, I wasn’t able to see that my assumptions often obstructed my insight into the girls’ actions and motivations.

Each aspect of the girls’ increasingly sophisticated social behavior relied on the development of cognitive and affective abilities. The ability to recognize and react to others’ feelings rested on growing awareness that others have similar emotions that can be influenced. This growth allowed them to be more compassionate. The awareness of others as distinct from oneself, each with a subjective mental state, made persuasion and negotiation more practicable. Greater language skills enabled them to influence others and let their play become more complex. Maturing abilities in self-monitoring and self-regulation allowed them to better control their impulsive, hurtful behavior and to adjust it when they found changed circumstances. It’s likely that their conflicted sparring helped them develop these capabilities.

Although less frequent and intense as the spring wore on, the trio’s emotional struggles with each other persisted. Children act to achieve goals, even if not consciously. The three of them achieved something by enacting these painful conflicts. Since children learn about themselves by encountering others, the skirmishes likely served to build crucial interpersonal knowledge—revealing how each girl felt, paired off, recombined, and made up. Conflicts are also chances to assert individuality, measure the strength of individual wishes, and compare divergent interests. The girls’ dramas of exclusion, abandonment, and ecstatic re-inclusion may have served as a kind of risky ritual that served to define their social network, control its territory, and test its durability.

There are three ways this study contributes to the broader field of early childhood. First, it illustrates how early childhood teacher research can illuminate our work with children (Meier & Henderson 2007). What began as a practical matter of dealing with intense conflict evolved into a purposeful, detailed study of the children’s behavior, practitioner methods, and the influence of conflict on the classroom. My “insider” status as their teacher enabled me to closely observe the girls’ behavior and the effects of our interventions on a day-to-day basis, something that an outside researcher could not do. The teacher research tools of systematic observation, thoughtful description derived from patterns of behavior, practitioner reflection, and cross-validation with colleagues enabled me to construct a deeper understanding of child behavior to inform my teaching and to offer to other teachers in similar situations.

Second, close examination of the girls’ interpersonal dynamics throughout the study also challenged my assumptions by helping reveal children’s surprising early competence in social interactions. It highlighted the intricacy of their thinking and planning about friendship, play, and belonging. It shed light on the purposes and social operations of networking and peer cultures among young children. These were revealed as fields of action for enhancing self-knowledge by experiencing the lives of other children; for building social skills; for establishing independence and autonomy from adults; and for extending learning through peer collaboration, as Vygotsky described the process. This all created changes in my teaching, and therefore in the girls’ interactions.

Third, by tracing how we helped the girls outline the social choreography of their small community and enhance their relational skills, this teacher research suggests strategies that other practitioners can try—and pitfalls to avoid—as they encourage children to manage conflict and choose from a broader range of prosocial approaches to classmates.

This study had the main effect of changing my thinking about peer interactions of all kinds, which improved both my thinking and my practice. My early response to the triad’s conflict focused on prevention; it was formulaic rather than thoughtful. With good intent, I sought to ease conflict by pressing for my adult notion of prosocial play—“enforced togetherness.” Like many teachers, I had assumed that conflict arises mainly from deficit: immaturity, or worse, antisocial impulses. Admonishing the girls to meet adult standards of behavior didn’t work. However, when I studied the intricate and intentional ways they interacted, I learned to see their actions as experimental efforts to carry out social and emotional goals.

We teachers began to provide more positive choices of prosocial action. We modeled examples of more engaging language. Via discussions about other children’s experiences and mental states, we equipped children to better perceive and identify each other’s feelings. Throughout, we respected children as architects of their own relationships: “Allowing children more space to develop their own authentic ways to interact can be more effective in strengthening collaborative play and building their social competence with peers” (Neimark 2012, 64). When we reinforced and supported the capabilities underlying prosocial relations, the girls could better choose to adjust their own behavior to the benefit of all three.

Play-based programs with a social development emphasis could benefit from trying similar strategies, as could any other kind of early childhood program. Other early childhood practitioners may experience conflict and injury far more challenging than our conflicted play triangle. Conflict is part of life. We must find ways to teach within and from conflict instead of just suppressing it. We should capitalize on children’s innate empathy and sensitivity, instead of acceding to the basically antisocial notion that humans are controlled by self-centered impulses. If we give children opportunities to build their strengths, we can reinforce their ability to get along with others throughout their schooling and lives and make a contribution to a more just and humane society in the process.